Comparison of statistics of controlled source tones and single ship noise in the deep ocean

Abstract

The deep ocean experiment, PhilSea09, was conducted April-May, 2009, in the central part of the northern Philippine Sea (22d N, 126d E). During one period in the experiment, the R/V Melville was station-keeping 35 km from the Distributed Vertical Line Array (DVLA) while seven tones, from 79 Hz to 535 Hz, were transmitted from a controlled source suspended below the ship. Recordings on the 1000-m section of the DVLA centered on the surface conjugate depth at 5026 m were dominated by the noise of this ship except at the controlled source tone frequencies. Using non-parametric statistical tests, the statistics of the spectral envelope at the tone frequencies are compared to the statistics of those for the nearby ship-noise-dominated frequency bins. When a high tone-signal to ship-noise ratio exists, the statistics of the tones differ from those of the ship noise at the 5% level of significance by the Kolmogorov-Smirnov two-sample test. Tone statistics are seen to be Gaussian distributed at frequency bands of low tone-signal to ship-noise ratios, whereas at high signal to noise ratios, the controlled source tones are non-Gaussian. When the signal to noise ratio is high, the statistics of the tone and of the noise are from different distributions. Both the tones and shipping noise travel approximately the same path to the DVLA, so these differences in the received field statistics represent differences in the statistical properties of these two acoustic sources themselves, not of the environment.